EEG Based Schizophrenia and Bipolar Disorder Classification by Means of Deep Learning Methods

Abstract

In this paper, different techniques based on deep learning algorithms used for the classification and diagnosis of patients with mental disorders i.e., schizophrenia and bipolar disorder, are presented. To this aim, the signals obtained from 32 unipolar electrodes of non-invasive electroencephalogram analysis are studied to obtain its main features. More specifically, the analysis performed utilizes an innovative radial basis function neural network based on fuzzy means algorithm. Furthermore, the analysis of the variance of statistical parameters and entropy is applied. In total, 312 subjects with schizophrenia and 105 patients with bipolar disorder have been evaluated. The results obtained show a correct classification in patients compared to healthy controls. The proposed methods achieved a better performance than other machine learning techniques such as support vector machine or k-nearest neighbour, with an accuracy close to 96%. It can be concluded that this type of classifications will allow the training of algorithms that can be used to identify and classify different mental disorders with very high accuracy.

EEG-Based Classification of Schizophrenia and Bipolar Disorder with the Fuzzy Method

“…According to research, bipolar disease patients can be correctly identified from recorded EEG data [17]. The Balanced Accuracy, Recall, Precision, and F1 values, as well as the suggested approach for schizophrenia and bipolar disorder [18], are shown in Tables 1 and 2, respectively. Support vector machine (SVM) and K-nearest neighbor (KNN) are well-known classification algorithms [19], [20], [21].…”

Automatic Bipolar Disorder Assessment Using Machine Learning With Smartphone-Based Digital Phenotyping

“…These data types include self-reports [14]- [17], sleep patterns [18], [19], recorded patient voices [20]- [22], location information (GPS) [24], [25], communication records [25], [26], and data from wearable devices [27], [28]. Physiological signals include heart rate variability (HRV) [29], [30], electroencephalogram (EEG) [31], [32], electrodermal activity (EDA) signal [33], and functional magnetic resonance imaging (fMRI) [34], [35]. In addition, patients' activities in the community, such as text posts and shared videos, have been analyzed to examine emotional changes [36], [37].…”